We address the problem of estimating the population density from mobile phone data. After critically examining the relevant network-based data sources (CDR, VLR and passive monitoring systems) a novel methodology is presented with two key novelties. First, it enables the fusion of cell-level and Location Area-level data from heterogeneous data sources. Second, it considers a novel tessellation scheme based on cell coverage maps, instead of cell tower locations. Furthermore, it allows to integrate data from different network operators onto a common reference grid. Within the proposed framework, a Maximum-Likelihood formulation for the population density estimation problem is developed and tested via simulations,showing a significant gain over existing Voronoi-based schemes.
- Ehemaliges Research Field - Mobility Systems